Age | Commit message (Collapse) | Author | Files | Lines |
|
Because the service running flags of Gnocchi and Aodh
are overwritten in the loop of the compute node, once the flag
become false, the tests on all subsequent nodes are skipped.
This change makes to use new flags for each compute node.
Change-Id: Idc807419151b34b29f0117c9a7708a16a20d70dd
Signed-off-by: Toshiaki Takahashi <takahashi.tsc@ncos.nec.co.jp>
|
|
|
|
* Because the type of the variable res[3] is bool,
we should not compare it with a string.
* The expression "or 'None'" is always the same result,
because 'None' is a constant value.
It should be included in the comparison.
Change-Id: I25defad98d612d21027218438ca2a28744292cf5
Signed-off-by: Toshiaki Takahashi <takahashi.tsc@ncos.nec.co.jp>
|
|
Change-Id: I3a706b9512c270ecb1340db93e2bb4baacab1475
Signed-off-by: Aric Gardner <agardner@linuxfoundation.org>
|
|
|
|
|
|
This change is to enable write_redis plugin in collectd
due to DMA local-agent is going to use the plugin.
Change-Id: I522b41278bb020a48f33ac18f4ad08351e51ca36
Signed-off-by: Tomofumi Hayashi <tohayash@redhat.com>
|
|
|
|
JIRA: BAROMETER-103
The lvm plugin will not build due to a deprecated library. This
patch will disable the building of the plugin until upstream collectd
can correct the issue. Collectd issue #2820
https://github.com/collectd/collectd/issues/2820
Change-Id: Idcc8e274cef87e1ad0c62cbe9097a24b298adc5f
Signed-off-by: TsaLaGi <aasmith@redhat.com>
|
|
Collectd loads config files in alphabetical order.
Ansible creates the config files for collectd.
In this patch we change the name ipmi.conf to 0_ipmi.conf
This will allow ipmi to be loaded before PMU.
If PMU is loaded before ipmi the container will crash.
This is a bug in collectd 5.8
Change-Id: I3334a3a1970f596e36f665698649d6c04fd38fbb
Signed-off-by: John O Loughlin <john.oloughlin@intel.com>
|
|
Remove --net=host as it should not be used when
port mapping with the -p option
Change-Id: I7d6d73fe176362fc74df99200a6d68d2a3ec0994
Signed-off-by: John O Loughlin <john.oloughlin@intel.com>
|
|
Update the userguide for intel_pmu with new "Cores"
option available in plugin configuration.
Added sample config with default configuration.
Change-Id: Ibe4d451131dc2c9ae0282f018c7e23596f4bf2c1
Signed-off-by: Kamil Wiatrowski <kamilx.wiatrowski@intel.com>
|
|
|
|
Change-Id: Iec6300fbae08a9d924bf9db8b9b9702bc568844b
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Updates the following doc with barometer-collectd container support
as well as Grafana and InfluxDB support in Compass4nfv installer:
docs/release/configguide/postinstall.rst
Change-Id: Id2c7e1167b0fd99f9200510de7caaa7b25052a5f
Signed-off-by: Ramamani Yeleswarapu <ramamani.yeleswarapu@intel.com>
|
|
|
|
|
|
|
|
Change-Id: I4a610d8686d758913d73ca684db41616daf391bd
Signed-off-by: John O Loughlin <john.oloughlin@intel.com>
|
|
Change-Id: I926d0d03781a98e3509d7f9c4f4eaffc1320919b
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
Since Ceilometer support was removed from the
collectd-ceilometer-plugin[1], the repo name is misleading.
So the repo name needed to be updated [2]
This change updates references to the collectd-ceilometer-plugin
repo to point to collectd-openstack-plugins repo:
* in deployment operations
* in documentation i.e. links to docs
Dependencies:
[1] https://review.openstack.org/#/c/490823/
[2] https://review.openstack.org/#/c/500768/
This is dependant on the repo rename
Change-Id: I43facbd53bca5df3f7e51d978ae2843c187282d4
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
|
|
|
|
* Skip RDT tests if the node is virtual.
* Push fake positive results to try avoid failure
TODO: Get the test passing.
Currently, the individual tests show a pass, however, the overall test shows a failure.
Change-Id: Iedfb125a15ed1b9d41a3d48148a4ac363a89b90b
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
* Update the check for timestamp to use state_timestamp instead
state_timestamp indicates the last change in state of the alarm,
timestamp is the last update to the alarm settings. This is not
what we're looking for with aodh tests, so check the state update
instead.
* Trigger an alarm state change before checking whether the alram
was updated. Notification in collectd aren't necessarily updated
at a regular interval, so we have to do some action and check whether
the alarm is updated.
JIRA: BAROMETER-91
Change-Id: I26d9bb1ec7e50e07053e69068d241e4c4cbc30dc
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
|
|
|
|
|
|
|
|
Change-Id: I0ae30a46d11d53cdc218e38d0b3a852aa5be147d
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
The SendNotification option in ovs_events is set to
false in Apex templates. This means that no notifications
will be emitted on link status change.
This patch updates ovs_events.conf.erb to match the
config in the samples folder, adn the ones used by
the reference container
Change-Id: Ia034b81aedc67c2ced51cbf831baca176218d7d3
JIRA: BAROMETER-91
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
|
|
Since the addition of DEFAULT_ARCHIVE_POLICY to
collectd-openstack, and support in Apex, the archive
policy used by Gnocchi for Barometer metrics is set
to "high" by default. This removes the need to update
the archive policy by hand for verification.
Change-Id: I074b4cad4926bd0e59ca04f915865ed52d0b6979
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
Change-Id: Iffba385bd8e565459288f95eadf8a52bb6cc0084
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
Change-Id: Ie146e0ac9ad5e8bc1721b51e4e5112db51462854
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
* MCE log plugin was failing becaue it wasn't enables
* Updated a check for kernel version to match Fraser
compute nodes on Apex
* RDT was failing when virtual compute nodes were used
* Add check to detect whether compute node is a VM
* Skip RDT when running in a VM
* Gnocchi metrics were failing
* increase the wait time between checking metrics
* CSV checks were failing on Hugepages
* functest was assuming that tests were running on
a system with at least two NUMA nodes
* Only check for metrics on node0 of the system
* Report was showing FAIL even if a test was not executed
* Update the report output to show SKIP instead
JIRA: BAROMETER-91
Change-Id: I4f70fc84fce80054a7aca2116357b0233f355f70
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
* Configure the DEFAULT_ARCHIVE_POLICY option.
* This will set the default archive policy to "high".
Change-Id: I6dd88da0dc0eec96c1c472ae97c0db2395a3fb0e
Signed-off-by: Emma Foley <emma.l.foley@intel.com>
|
|
By merging several of the RUN lines and executing a yum cleanup all
at the end of the dependency installation, the resulting container
image will be around 300MB smaller.
Change-Id: I63139a5283ea40ec8ea5cc3fd7aa7faddc90f3ce
Signed-off-by: Leif Madsen <lmadsen@redhat.com>
|
|
Follow on to the change #52775
Fixed the if/else logic while installing the mcelog. And
added log messages to show on which node, mcelog is installed.
Change-Id: If5f61b1a780abd243516699e352a55baf0db96f7
Signed-off-by: Sharada Shiddibhavi <sharada.shiddibhavi@intel.com>
|
|
|
|
|
|
|
|
Change-Id: I342016af10a1ed3a11364bda7bd7a563882fafbd
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
|
|
Change-Id: I6e7cb3e0923495e236b74444b1da3f4450122d02
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
This file was renamed to avoid linking Kubernetes testcases to
OpenStack credentials.
This patch is mandatory for gating.
Change-Id: I6b4e7e83d6dfc8cf1d60bf016ef258a273156ffe
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
It inherits from xtesting.core.feature [1].
[1] http://xtesting.readthedocs.io/en/latest/apidoc/xtesting.core.feature.html
Change-Id: Ib547aa60d9b8bf61ab5433496d7149a43355d08c
Signed-off-by: Cédric Ollivier <cedric.ollivier@orange.com>
|
|
Change-Id: Ida03c47bbe94600b034c95bd0d795e62c253303c
Signed-off-by: Gordon Kelly <gordon.kelly@intel.com>
|
|
|
|
|